Quote:
Originally Posted by shaunpw
ok, here we are with todays update.
I have now replaced IC1 & 2 (cd4066), IC24, 25, 26, 27 (all now OPA2134)
All the tests carried out with an input short plug.
I got myself a better DVM, now reading proper millivolts.
So, with radio selected, no resistors shorted, I get -41.1mv/-37.mv on IC27 op.
as above with cd selected, I get -4.5mv/-3.7mv on IC27 op.
|
The only difference in signal path between radio and CD inputs is the internal switch selection within the 4066. Although all should be 0.00 volts you have a ten fold difference between the two here. There is however one important difference...
Look at the circuit.
If C89 and C90 are present and good then we can discount anything on the input side of those caps. The important difference I mentioned are the values of R143 and R144 (10k) and R11 and R13 (100k).
Your fault at this stage seems to suggest to me that the issue is caused by leakage somewhere. The 10k effectively pulls those nodes down better toward zero volts than 100k and so reduces the apparent DC voltage.
Lets try and be logical with this, the offset has a negative polarity at IC27 output and the opamp is configured for unity gain so that implies the leakage is from a negative rail onto the signal line.
I've drawn the signal line in red for one channel. I suspect anywhere on this line is seeing leakage from print in an area that runs close to a negative rail.
A quick off the wall calculation... -/+8 volt rails (so we work with 8 volt) means we have a 'leakage' resistance of around 18 meg ohms. That would cause the 100k to see 45mv across it. Swap that to 10k and we get 4.5mv.
I think something like this fits your fault symptoms. What to do though... I'm a great believer in washing boards but I'm worried here whether any leakage is down more to something with the board material or soakage into it.
Are they fibreglass boards or paxolin?
Given that modern signal sources (not old DIN standard) can happily drive loads down to about 600 ohm you could reduce the 100k's (and the 10k's).If you remove (link out) C89 and C90 then with FET opamps fitted the output of IC 24 should force a zero volt condition... assuming there is no similar leakage at that opamp input.
If IC 24 has 0.00 volts at its output now then you could do that trick. If not the problem is essentially confirmed as leakage across the whole board (imo), however R135 and R136 could be taken much lower if that input was used with a CD player and that would again force a near zero volt condition.